They are trying to predict what types of html pages cause parsing problems. They have a set of variables that describe an html page. They are trying to figure o

They are trying to predict what types of html pages cause parsing problems. They have a set of variables that describe an html page. They are trying to figure out which variable or combination of variables predict a bad parsing outcome. Think of cluster analysis as simply plotting those variables as points on a graph. Then drawing a circle around points that are close to each other.